Application of Two Smart Nonlinear Multivariate Calibration Methods for the Concurrent Quantitative Spectrophotometric Determination of Antiparkinson Drugs in Pharmaceutical Formulation and Biological Samples: Comparison with HPLC

Abstract

Chemometric-assisted UV-spectrophotometric methods, including least squares support vector machine (LS-SVM) and partial least squares (PLS) as multivariate approaches, were proposed for the quantitative simultaneous determination of levodopa (LEV) and benserazide (BS) in pharmaceutical formulation and urine samples. In the LS-SVM method, the related parameters, named the regularization parameter (γ) and width of the function (σ), were optimized, and the values with the minimum root mean square error (RMSE) were selected. The RMSE values were obtained at 0.9246 and 0.3423 for LEV and BS, respectively, whereas in the PLS model, the RMSE of the test set was found to be 0.3674 and 0.1216 for LEV and BS, respectively. The suggested models disclosed satisfactory recovery related to the synthetic mixtures in the range from 91.21 to 107.90% for LS-SVM and from 94.33 to 101.42% for PLS. The simultaneous determination of the LEV and BS in tablet dosage form and spiked urine samples using the proposed models revealed recovery higher than 94% and 91%, respectively. A comparison was made with the ANOVA test between the proposed methods and highperformance liquid chromatography (HPLC), and no significant difference was shown. These chemometrics methods are fast, facile, inexpensive, precise, and do not require sample pretreatment. Low solvent use, reduced energy consumption, and short time for analysis are other advantages of these methods. Therefore, they can be a safe and stable approach for drug analysis in quality control laboratories instead of expensive and time-consuming chromatographic techniques.